A.I.D. PROGRAM DESIGN AND EVALUATION METHODOLOGY REPORT NO. 8 (Document Order No. PN-AAL-088)

Transcription

1 CONDUCTING GROUP INTERVIEWS IN DEVELOPING COUNTRIES A.I.D. PROGRAM DESIGN AND EVALUATION METHODOLOGY REPORT NO. 8 (Document Order No. PN-AAL-088) by Krishna Kumar (Bureau for Program and Policy Coordination, A.I.D.) U.S. Agency for International Development April 1987 The views and interpretations expressed in this report are those of the author and should not be attributed to the Agency for International Development. TABLE OF CONTENTS Foreword Acknowledgments Summary 1. Introduction 1.l Description of Group Interviews 1.2 The Use of Group Interviews by Project and Program Managers 1.3 Advantages and Limitations of Group Interviews 1.4 Time, Staff Requirements, and Costs 2. Planning Studies Using Group Interviews 2.1 Delineating the Scope and Objective 2.2 Reviewing Available Information 2.3 Defining the Key Concepts 2.4 Selecting the Appropriate Technique 3. Conducting Focus Group Interviews 3.1 Preparing the Interview Guide 3.2 Size and Composition of the Group 3.3 Sampling 3.4 Location and Seating Arrangements 3.5 Timing and Duration

2 3.6 Opening the Interview 3.7 Probing Techniques 3.8 Pacing the Discussion 3.9 Controlling Discussions 3.10 Controlling Group Pressure 3.11 Recording Discussions 4. Conducting Community Interviews 4.1 Structured Interview Guide 4.2 Selecting Communities 4.3 Size and Timing of the Meeting 4.4 Interview Team and Protocols 4.5 Balancing Participation 4.6 Generating Quantitative Data 4.7 Informal, Post-Meeting Conversations 4.8 Recording Community Interviews 5. Group Interviews and Interviewer/Moderator Biases 5.1 Hypothesis-Confirmation Bias 5.2 Consistency Bias 5.3 Elite Bias 5.4 Concreteness Bias 5.5 Increasing Interviewer Awareness of Potential Bias 6. Suggestions for Further Reading Bibliography FOREWORD There has been a growing belief among A.I.D. managers that many widely used data collection methods, particularly censuses, sample surveys, and detailed ethnographic descriptions, are often not appropriate for generating information for decision-making. They require considerable investment of resources, take a long time to complete, and often produce data that are too elaborate for their intended purpose. Researchers, therefore, have begun to use alternative methods that can provide timely information in a cost-effective manner. The group interview is one of them. Group interviews can be used for obtaining a wide range of information for different purposes. They can provide background information for designing projects and programs, generate ideas and hypotheses for intervention models, provide feedback from beneficiaries, and help in assessing responses to recommended innovations. They are also useful for obtaining data for monitoring and evaluation purposes and for interpreting data that are already available within A.I.D. Despite their immense potential, the literature on group interviews is extremely limited. The little material that exists deals with group interviews within the environment of

3 industrialized nations. Therefore, the Center For Development Information and Evaluation of The Bureau for Program and Policy Coordination has produced this guide, which describes the nature, uses, advantages, and limitations of group interviews with reference to the conditions of developing countries. It also explains the steps involved in conducting two types of group interviews -- focus group interviews and community interviews. The guide has been written in jargon-free language for use by A.I.D. managers, contractors, and host country officials. I am sure that they will find it useful. Haven W. North Associate Assistant Administrator Center for Development Information and Evaluation Bureau for Program and Policy Coordination Agency for International Development ACKNOWLEDGMENTS The author has received valuable comments from his colleagues, particularly Paula Goddard, Annette Binnendijk, Chris Hermann, Nena Vreeland, Ned Greeley, Maureen Norton, Pamela Johnson, Carolyn Weiskirch, Cindy Clapp-Wincek, and Ed Dela Rosa. In addition, Josette Murphy, of the World Bank, gave valuable suggestions. The author wishes to acknowledge his sincere thanks to all these friends. Above all, he is indebted to Mr. Haven North for his overall interest and encouragement. SUMMARY The Group interview is one of the rapid, cost-effective data collection methods. It involves the use of direct probing techniques to gather information from several individuals in a group situation. Although superficially the difference between the individual and group interviews is the number of participants, this difference contributes to major variations between the two with regard to planning, nature of interview guides, probing techniques, and analysis of information. Group interviews can serve a wide range of information collection purposes. They can provide background information and help to generate ideas and hypothesis for project and program design, provide feedback from beneficiaries, and help in assessing responses to recommended innovations. They are also useful for obtaining data for monitoring and evaluation purposes and for interpreting available quantitative data. There are two main types of group interviews -- focus group interviews and community interviews -- that have wide potential in developing countries. Both types should be carefully planned. The investigator should conduct a systematic review of the

4 relevant documents, records, or studies and consult with a few key informants before venturing into the field. The main concepts should be clearly defined in order to avoid possible misunderstanding between the respondents and the interviewers. Focus Group Interviews The focus group interview represents a group situation in which the participants talk with one another under the guidance of a moderator for the purpose of generating relevant ideas and information. The main interaction is among the participants, who stimulate each other. Such interviews are conducted on the basis of a flexible interview guide. The interview guide lists the main subtopics and issues to be covered. The number of items in the guide should not exceed six or seven, in order to leave enough time for in-depth discussions. The optimal number of participants ranges between six and ten. A group of this size is neither too large nor too small and can be easily managed. The composition of the group should be homogeneous, with members sharing similar background and experience. Although probability sampling cannot be used for selecting the participants, every effort should be made to ensure that they are representative of the population in which the investigator is interested. Seating arrangements should facilitate maximum interaction among participants. The best arrangement is to have the participants sit around a table facing each other. The duration of a session can range from 1 to 2 hours. However, if the discussions are interesting, the group can continue beyond the scheduled time. After a brief introduction, the moderator should explain the purpose of the inquiry. The moderator should stress that it is an informal group and that participants should express their views candidly. Probing should be subtle, and the moderator should adopt a posture of "sophisticated naivete" to encourage the participants. For eliciting details, questions should be asked casually. The queries of what, when, where, which, how, and why are quite helpful. Visual aids such as slides, films, and pictures can be shown to stimulate discussions on a specific subtopic. To prevent a few individuals from dominating the discussions, the moderator can follow one of three strategies: (1) give nonverbal cues to the respondent to stop, such as looking in another direction, showing a lack of interest, and stopping note taking; (2) politely intervene, saying that he has somehow missed the point and would like to summarize what the respondent was saying -- then refocus the discussions; (3) take advantage of a pause and say that the issues raised are of vital significance and should be discussed in a separate session. To minimize group pressure, which inhibits dissenting participants from expressing their views, the moderator can ask

5 for other ideas, explanations or recommendations than those already discussed, suggest new ideas for discussion. The moderator can also look at the participants who appear to be skeptical of the views of the group and encourage them to speak. Focus group interviews should be carefully recorded. The notes should include (1) records of the discussions, (2) moderator's observations of the nonverbal behavior of the participants, and (3) the moderator's own ideas, hunches, or thoughts generated during the interview. Community Interviews Unlike the focus group in which participants discuss a subject among themselves, in community interviews the primary interaction is between the interviewer and the participants. Community interviews take the form of community/village meetings open to all adults in the area. Communities for conducting interviews are carefully selected to minimize sampling biases. A structured interview guide listing precise questions for the interview should be prepared for community interviews. Its language should be simple. Both leading questions and questions that combine two or more queries should be avoided. Questions on controversial topics that can generate strong emotions among the participants should also be avoided. The interviewer has little control over the the size of a community meeting, which depends on factors such as the population of the community, time and location of the meeting, and interest of the target population. Past experience has shown, however, that if they are well publicized and the time is convenient, many people turn up. If the number of participants exceeds 30, the investigator should consider dividing the group into two or more subgroups for more open discussions. As far as possible, community interviews should be conducted by a team of two because it is extremely taxing for a single interviewer to preside over a large meeting, ask questions, probe respondents, and take extensive notes. Restraining the leaders from monopolizing the interviews requires great acumen and interpersonal skills on the part of the interviewer. One strategy that has proved effective in the past is to discuss the main subtopics with important community leaders before the meeting. This strategy has two merits. First, the leaders might not want to repeat themselves in large meetings once they have articulated their views and shared their information. Second, the interviewers can say at the outset of the meeting that they have already discussed the subject with a few leaders and have come to hear from the others. Community meetings can be also used for generating some quantitative data. Often, useful community-level statistics can be gathered from them. Experience has shown that such data are relatively accurate. An interviewer can also take polls during

6 community meetings for obtaining quantitative information about the behavior, experience, opinions, and assessments of the respondents. Extreme caution is required for making generalizations on the basis of such data. Post-meeting conversations with interested individuals are an integral part of the interview process. They enable the interviewer to get the views of those participants who for some reason preferred not to express themselves in the meeting. Interviewers' Biases There are four kinds of interviewer biases in group interviews that can undermine the validity and reliability of the findings. The "hypothesis-confirmation bias" arises from selectively focusing on information and ideas that confirm the preconceived hypotheses and beliefs of the interviewers. The "elite bias" results from an interviewer's tendency to give more weight to the views of elites than those of others. Vivid descriptions and statistical data about a few cases can give the impression that they represent general situations, thereby producing "concreteness bias." Finally "consistency bias" can be generated by a premature search for coherence in the disparate, often irreconcilable remarks of the participants in order to draw meaningful conclusions. A.I.D. managers should be mindful of such biases in the studies contracted by them. 1. INTRODUCTION This monograph describes the nature and mode of conducting group interviews in developing countries.{1} It attempts to explain in nontechnical language their nature, types, uses, and limitations. More important, it details the steps involved in conducting two types of group interviews with great potential for application in developing countries: focus group interviews and community interviews. The following subsections of Section 1 describe group interviews and their primary uses and advantages. Section 2 presents the general steps involved in planning studies based on group interviews. Section 3 highlights focus group interviews and discusses the techniques for conducting them, and Section 4 outlines the procedures for conducting community interviews. Section 5 describes biases of interviewers and moderators and their adverse affects on the interview process. Section 6 presents a brief discussion of suggested reading. {1}There is almost no literature on group interview techniques, especially on conducting them in developing countries. This brief monograph attempts to provide some guidelines for researchers who want to use group interview techniques in developing countries but have no sources to draw on for guidance. 1.1 Description of Group Interviews

7 Broadly speaking, group interviews involve the use of direct probing techniques to gather information from several individuals in a group setting. Such interviews can be conducted by one or more interviewers, with or without an interview guide, and with groups of varying sizes and composition. Superficially, the difference between individual and group interviews is the number of respondents; however, this difference leads to major variations between them with regard to planning, nature of interview guides used, probing techniques, and analysis of information. The two types of group interviews that will be discussed are focus group interviews and community interviews. Focus group interviews focus on a specific topic and are conducted in small group sessions. A distinguishing feature is that participants discuss ideas, issues, insights, and experiences among themselves. Each member is free to comment, criticize, or elaborate on the views expressed by previous speakers. The moderator guides the discussions toward the issues identified in the interview guide and uses various probing techniques to elicit further information. Focus group interviews can last 1 to 2 hours, depending on the interest of the participants and the subject under investigation. The groups generally have 6 to l0 participants, selected on the basis of a set of criteria dictated by the objective of the inquiry. Community interviews are also based on an interview guide and take the form of public meetings that are open to all members of a community or village. Usually, the large number of participants (more than 15) does not permit free discussions among all those present. Each participant is not expected to answer all the questions raised in a session. Generally, a few leaders tend to dominate the discussions. Use of an interdisciplinary team rather than a single interviewer is more effective in conducting community interviews. This monograph is written primarily to meet the needs of practitioners who are involved in the design, monitoring, and evaluation of development projects. Often they require data and information of reasonable accuracy within a limited time span. What they need is timely information to guide their activities. For them, group interviews could be very useful in gathering relevant information. 1.2 The Use of Group Interviews by Project and Program Managers The managers of A.I.D. projects and programs and host country officials can use group interviews for generating a variety of information on development interventions. Specifically, they can use them for the following purposes.

8 1. Securing background information for project and program planning. Group interviews can provide much needed background information about communities and villages, social and economic systems, farming systems, target groups, and local organizations and institutions. For example, if A.I.D. is designing an agricultural extension project, a series of group interviews with the intended beneficiaries, concerned officials, and other experts can provide relevant information about farmers, agricultural land, farm practices, crops, marketing facilities, and the outreach capacity of existing extension services. 2. Generating ideas and hypotheses for project and program design. Group interviews are ideal for exploratory investigations. For example, at the project or program design stage, they can be conducted for generating ideas and hypotheses concerning the needs and requirements of target groups, the appropriateness of the intervention models, and the suitability of delivery systems. Obviously, these hypotheses will have to be tested on the basis of additional data gathered from other sources. 3. Getting feedback from project/program beneficiaries. The majority of development projects build or improve delivery systems for providing necessary goods and services to the target populations. Hence the managers need constant feedback from their beneficiaries about the efficiency and effectiveness of such systems. For example, the managers of an agricultural extension project should know whether extension workers are disseminating the message in a language farmers understand, whether they hold agricultural demonstrations at convenient locations and times, and whether the message imparted is relevant and timely. Such information can be rapidly obtained through community interviews or focus group discussions with the targeted farmers in select locations. 4. Assessing responses to recommended innovations. Nearly all development interventions introduce innovations, and the ultimate success of the intervention depends on the acceptance of the innovations by the targeted group -- farmers, small businessmen, or government officials. Group interviews can be helpful for assessing the reactions of the intended clients to the innovations. For example, focus group interviews can inform the project staff in a family planning intervention about the reactions, concerns, and reservations of women about the use of contraceptives. Focus group interviews can often reveal more in-depth information than can formal sample surveys. 5. Interpreting available quantitative data. Donor agencies, host governments, and project staff gather data from various sources. Group interviews can be helpful in interpreting them; they can add a qualitative dimension, which the data lack. 6. Investigating implementation problems. Many implementation problems faced by development projects require in-depth probing of the motives, attitudes, and understanding of the actions involved. It is important to know, for example, why

9 the majority of farmers in a project enthusiastically adopt a new variety of wheat but revert to their previous variety in succeeding seasons, or why demand for contraceptives soars in some regions but remains stagnant in another, or why farmers are not repaying their short-term loans to the agricultural credit societies. In all these examples, we need to understand the perspective of the individuals -- why they are behaving as they are. Group interviews, particularly focus group interviews, can be very useful for this purpose. 7. Project onitoring. Group interviews can also be helpful for generating information needed to monitor project performance, achievements, and limitations. They can also generate practical recommendations and suggestions for improving project performance. 8. Evaluation. Finally group interviews could be a major source of data for mid-term, terminal, and impact evaluations. 1.3 Advantages and Limitations of Group Interviews As a mode of data collection, group interviews have several advantages. First, they enable the investigator to gather information rapidly. A group interview with 6 to l0 people can be conducted within an hour or two, which is much less time than it would take to interview them individually. While it will not provide the same depth of information that might be gained in individual interviews, a skilled interviewer can still obtain considerable information and understanding from group interviews. When impending deadlines impose severe time constraints, group interviews may be the only viable alternatives. Second, group interviews are economical compared with in-depth interviews or structured surveys because they do not require a large staff of enumerators. A trained interviewer can conduct two or three group interviews in a day, each covering multiple respondents. Thus costs are considerably less for studies based on group interviews. Third, group participation can sometimes reduce individual inhibitions, thereby providing information that might not be otherwise shared. In some cases, people in groups are willing to share feelings, emotions, and concerns that they would be reluctant to express in more private settings. The obvious reason is that they find a sense of security in the group, which is undoubtedly an important consideration in rural conditions, where respondents are uncomfortable in the presence of outsiders. The knowledge that other farmers have the same reservations about the recommended technical package can lead a cautious farmer to express his own doubts on the subject in the presence of senior officials. Of course, the reverse is also true; some may be reluctant to express their intimate feelings and opinions in public.

10 Fourth, respondents are able to raise issues and concerns that the investigator might not have considered. Group members generate new ideas and approaches because they stimulate each other. A group dynamics emerges that encourages the participants to respond to each other's ideas and comments, thereby opening up fresh lines of inquiry. For example, farmers may remark casually that since the establishment of a procurement center by the government, they have had to transport their produce to the center themselves, instead of using private traders who used to pick it up at the farmgate. This comment may have a "snowballing" effect in starting an interesting discussion of several issues that might not have occurred to the investigator during the design of the investigation. In addition, group interviews can reveal divergent perspectives and innovative ideas, which are usually not identified in a structured survey. Fifth, one of the major advantages of group interviews is that they permit a direct interaction between respondents and the investigator. In this respect, they are better than formal survey interviews conducted by enumerators who are not involved in data analysis and interpretation. Group interviews provide investigators with a broad view of the situation; they are able to listen to respondents and also watch their expressions. Sixth, information gathered in group interviews can sometimes be more accurate than that obtained in individual interviews because respondents are generally reluctant to give inaccurate answers for fear of being exposed by other participants; when they do, others tend to correct them. In the oft-cited example of group interviews conducted by Ladejinsky in Bihar, India, a large landowner who said that he owned only 30 acres of land (the maximum permitted under law) grudgingly conceded that he owned 300 acres when the other participants humorously questioned the validity of his answers (Ladejinsky 1969). Finally, group interviews provide considerable flexibility to the interviewers, who are not unnecessarily constrained by their research instruments and are thus able to pursue the leads provided by respondents. The advantages of group interviews should not, however, obscure their major limitations. First, group interviews cannot generate reliable quantitative data from which generalizations can be derived concerning the whole population. They give a relatively accurate picture of the prevalence of a phenomenon, attitude, perception, or behavior pattern, but not of its extent or pervasiveness. For example, an investigator may learn from group interviews that small farmers are not availing themselves of the short-term agricultural credit offered by the project institutions because of the cumbersome delay in processing loan applications. But the investigator could never know what percentage of farmers in the project area are being deterred by this factor. Hence, when precise quantitative data are required, group interviews cannot serve the

11 purpose. Even when some limited quantitative information is generated in community interviews, the interviewers should resist the temptation to generalize from the findings. Second, group interviews are highly susceptible to interviewer biases, which can undermine the validity and reliability of their findings. Probably the most common bias is "hypothesis confirmation bias." Unless interviewers are well trained and possess a healthy skepticism about their own hypotheses and hunches, they may misinterpret the group discussions as confirming their own position. Experience has shown that interviewers can consciously or unconsciously give greater weight to the views expressed by elites than those of other members of a group, thereby projecting a distorted picture. (Four types of interviewer biases are discussed in Section 5.) Finally, particpants do not divulge sensitive information in group situations. 1.4 Time, Staff Requirements, and Costs A field study based on 10 to 15 group interview sessions should be completed within 6 weeks under normal conditions. Usually it takes about 1 or 2 weeks to review literature and to develop the interview guide, 2 to 3 weeks in the field to conduct the interviews, and another 2 weeks to prepare the report and recommendations. This estimate assumes that the investigation team will be able to complete at least one group session every day while in the field. Thus, if the geographical area to be covered is large, and there are problems in moving from one site to another, additional time will be required. The requirements for the research staff are different for focus group interviews and community interviews. Focus group interviews are conducted by a single moderator, whereas community interviews require a team of two. When the principal investigator is not fluent in the local language, he or she must be assisted by a local collaborator. Investigators should ideally meet three requirements. First, they should have appropriate training in conductiong group interviews. A continual difficulty faced by A.I.D. managers is that many people claiming to be experts have little experience and expertise. Second, the investigator must have substantive knowledge of the subject under investigation in order to understand and correctly interpret the discussions and responses of the participants. For example, only a person acquainted with extension services should interview farmers on this subject. Finally, the interviewer should be sensitive to the cultural norms of the community. The major costs of group interviews are the remunerations, per diem, and traveling expenses of the investigator(s). These expenses can be quite low for local experts from the host

12 countries. For focus group interviews, additional costs might be incurred for renting rooms for the discussions and for modest honoraria for the participants (honoraria should be paid only when absolutely necessary). In addition, secretarial services should be considered when estimating costs. 2. PLANNING STUDIES USING GROUP INTERVIEWS This section briefly describes how to plan studies based on group interviews. The steps involved are similar to those for other investigations requiring empirical data collection: (1) delineating the scope and objective of the inquiry, (2) reviewing the available information, (3) defining the key concepts to be used, and (4) selecting the appropriate group interview technique. The interviewer should follow these steps before venturing into the field. 2.1 Delineating the Scope and Objective One simple rule in survey research is that the investigators should look backward; they should focus on the ultimate, desired outcomes and then examine how these will be achieved, if at all, by a survey. In other words, the focus should be on the questions that the survey is intended to answer. This rule is as valid for group interviews as it is for formal surveys. Thus the first step is to prepare a set of questions, which the A.I.D. manager would like to be answered by the study. This can be illustrated with a simple example. Consider a project that has been supplying a set of inputs to farmers for growing an improved variety of maize. Despite the best efforts of the management, the targeted farmers have shown little interest in the recommended technical package. As a result, the USAID Mission has decided to initiate an exploratory study to gain an understanding of the farmers' perspective through group interviews. In this case, the investigator should begin by writing down research questions such as the following: 1. Do farmers regard the improved variety of maize as profitable, given the additional investment of time and resources? 2. Are interested farmers able to procure the required inputs such as seeds, fertilizers, insecticides, and pesticides in time? 3. Do they face problems in securing short-term credit from the project to purchase the required inputs? 4. Is there a shortage of labor during the peak agricultural seasons that constrains interested farmers from cultivating the improved variety of maize?

13 5. Is there a gender issue? Which farmers, male or female, are expected to do extra work? 6. Is the taste of the improved variety palatable to the local populace? 7. Do farmers believe that they are likely to get timely technical advice about the package from the agricultural extension service when needed? 8. What are the socioeconomic characteristics of farmers who are using the improved variety of maize? 9. What proportion of farmers abandoned the new variety after initial trial? 10. What proportion of targeted farmers is actually using the new variety? Once the list has been prepared, the investigator should examine whether these questions can be explored through group interviews. Some study questions cannot be explored in group interviews, and in such cases, the plan for group interviews will have to be abandoned. For instance, group interviews cannot answer questions 7, 8, and 9 because they require statistical data that can be collected only through formal surveys. The remaining questions, however, can be adequately dealt with on the basis of group interviews. It is prudent to discuss the study questions in detail with the concerned A.I.D. or host government staff. This is necessary because unlike the questionnaires prepared for formal surveys or censuses, interview guides used in group interviews do not provide a clear idea of the nature and contents of the interviews. As a result, there is potential for genuine misunderstanding; the expectations of the sponsors might be different from those of the researcher. 2.2 Reviewing Available Information The next step is to review the information available on the subject. This is necessary for clarifying the essential ideas and for avoiding duplication. Launching into group interviews without sufficient knowledge of the local conditions and the problem to be investigated is a waste of time. While the interviewers need not conduct a comprehensive library search as in an academic research effort, they should familiarize themselves with the literature and discuss the subject with a few key informants. The following are some important sources of information that can be rapidly tapped for this purpose:

14 -- Project/program records and documents. These are easily available and can provide a wealth of information. For the above-cited example of the failure of the target population to adopt the new technical package, the reports and records of extension services, of the project management, or of agencies providing the inputs can be extremely useful. The reports prepared by extension workers can tell the investigator about the initial enthusiasm, reservations, or problems of farmers, while the records of the agencies supplying inputs can shed light on the nature of the demand for the various inputs and their timely availability. The progress reports of the supervisory staff can give additional insights about the overall management of the project. -- Published or unpublished studies undertaken on the subject by different donor agencies, governments, and educational institutions. These can include feasibility studies, evaluations, doctoral theses, and formal and informal surveys. Often the existence of such studies is not known except to the sponsoring agencies and the researchers who participated in them. They can be located only through personal visits to the offices of concerned organizations and institutions. -- Secondary data available from government and research institutions. One is often surprised by the wealth of data routinely collected by the governments in developing countries that can be obtained with little effort. Such data are usually available from statistical offices, census bureaus, ministries of agriculture, health departments, and planning divisions. In addition, many private and public research firms also gather statistical data on a limited range of topics. In addition to these sources, the investigators can meet and discuss the subject with a few knowledgeable individuals. For instance, they can talk with extension staff to get information about farming systems, with public health workers to learn about family planning practices, and with officials of agencies supplying inputs to learn about the sales of fertilizers, seeds, or insecticides. 2.3 Defining the Key Concepts The conventional wisdom in qualitative research is that the relevant concepts should not be defined in advance but should evolve from the field experience. This advice, although sound for academic research, is not appropriate for studies based on group interviews unless their purpose is to refine or develop new concepts. The obvious reason is the short time in the field: group interviews are conducted within days rather than weeks. Moreover, the concepts used are not new and have generally been

15 used in previous investigations. The investigator should carefully define the key concepts during the planning stage. Conceptual clarification avoids confusion and helps to sharpen the focus of inquiry. Moreover, if the investigator does not clearly define the concepts, the investigator and the respondents might be talking about different things; while the investigator is referring to smallholders, respondents might have large commercial farmers in mind. As for as possible, the definitions should be simple and congruent with the common usage of the terms; otherwise the participants in group interviews might not comprehend their precise meaning. Box 1 describes the case of a group interview in Thailand, where the use of the word "marriage" created some problems during the interview. Only a few major concepts need to be defined for the purpose of the study. For example, in an investigation of farmers' responses to the new varieties of maize seed, one need only define what is meant by the "farmer" in this context, which improved varieties are involved, and what the technical package promoted by the project contains. One advantage of the group interview is that inappropriate definitions can be modified during the course of the study if misinterpretation becomes apparent. 2.4 Selecting the Appropriate Technique Finally, an appropriate group interview technique should be identified. The two group interview techniques most relevant in the context of developing countries are focus group interviews and community interviews. The most important consideration in selecting the specific group interview technique is the nature of the information required. When in-depth knowledge about peoples' perceptions, feelings, or values is sought, focus group interviews tend to be more useful. They permit an intensive probing of a subject and bring to the surface the subjective feelings, opinions, and Box 1. The Importance of Defining Key Concepts: Marriage in Thai Society In Thailand, couples may choose any one of several forms of conjugal union, from simply moving in together to having an elaborate ceremony. As a result, it is extremely important to state explicitly which meaning of the term is intended. This was highlighted in a group interview situation, as the following excerpts indicate: In Thai, the usual, more formal word for marriage in general also has a more specific and literal connotation of a marriage ceremony. While we

16 anticipated that the use of this word in our focus group sessions would prove no problem and be interpreted in its more general sense, a number of participants who had not apparently had a marriage ceremony obviously interpreted the term more literally than expected. The following quotes underscore the need for researchers to be aware of the variety of ways conjugal unions can be initiated in Thailand and the different terms used. "I didn't get married. Back then we just turned out the lamp and did it." (older woman, North) " I did not get married, I ran away [eloped]." (younger woman, construction site, Bangkok) "For those who are laborers, they will not get married but will run away together [elope]." (older woman, construction site, Bangkok) "Back then, you did not get married, not like nowadays." (older woman, North) Source: Pramuslratana, Havanon, and Knodel (l983, l4-15). judgments of participants. Thus they are very useful when the views of a specific target group or market segment are sought. Community interviews are most useful for understanding local needs, expectations, and behavior patterns. Suppose we want to know if communities will be willing to share the cost for constructing a school building or a health center; community interviews can give us a reasonable idea. They can also provide relatively accurate information about community-level indicators such as access to roads, availability of medical facilities, major crops grown, or marketing facilities. There are no hard and fast rules about the selection of the group interview technique; each investigator will have to exercise individual judgment in the matter. 3. CONDUCTING FOCUS GROUP INTERVIEWS Focus group interviews offer a means of getting in-depth information on a specific topic through a discussion group. The underlying premise is that people who share common experiences, problems, or concerns are willing to reveal them in a group atmosphere. Focus group interviews are not simply individual interviews conducted in a group setting; the moderator does not ask the same question of all respondents. Rather, focus group interviews represent a group situation in which the participants talk with each other under the guidance of a moderator. Each participant is stimulated by the comments of others and in turn stimulates them. Strictly speaking, when participants direct

17 their questions to the moderator rather than to each other, it is not a focus group interview. The primary role of the moderator is to promote group discussions. Interest in focus group interviews has been growing since the 1960s. They have been used extensively in marketing studies for understanding the psychological and behavioral underpinnings of consumer behavior and for identifying the ways and means to affect it. The breadth of topics on which focus group interviews have been conducted has been unusually large -- from testing a concept to gauging ultimate satisfaction with the product. Educators have also used focus group interviews for evaluating curricula and the performance of instructors -- reflecting an awareness of how focus group interviews can complement the information gathered through the formal questionnaires normally administered for this purpose. Public health professionals have used them for assessing the needs and perspectives of clients in health delivery systems. Finally, investigators working on family planning, especially the social marketing of contraceptives, have used focus group interviews to gather useful knowledge for their own purposes and have also demonstrated the effectiveness of focus group interviews in different social and cultural contexts. The following subsections outline the steps involved in conducting focus group interviews. (Planning group interviews was discussed in Section 1.) 3.1 Preparing the Interview Guide Once the focus and scope of an inquiry have been determined, the major concepts defined, and the target populations identified, an interview guide is prepared. The guide, similar to that used in in-depth qualitative interviews with individual respondents, lists the subtopics to be covered. The guide does not give detailed instructions to the moderator; rather, it is an aide memoire to refresh the memory during the interview. The number of items included in the guide should be limited. The moderator should focus on only six or seven subtopics, but should explore them in considerable detail. Having fewer items leaves more time to pursue leads before moving to another item. When an attempt is made to cover many subtopics in a single session, participants do not feel free to raise issues, questions, or ideas that are uppermost in their minds because they feel constrained by time. The nature of an interview guide can be illustrated with a concrete example. Suppose an investigator is conducting focus group interviews to uncover attitudes toward family planning and the use of contraceptives among the target population. The purpose is to gain a general understanding of the subject so that a comprehensive sample survey can be launched at a later stage. In this case, subtopics such as the following can be included in the

18 interview guide: -- Preferred size of the family -- Preferences concerning the sex of children -- Changing attitudes and behavior patterns related to age at marriage -- Economic aspects of having children (e.g., effect of family size on standards of living, importance of children in old age) -- Attitude towards childless couples; reasons? -- Attitude toward family planning -- Awareness of different contraceptive methods Many more subtopics than those identified in the interview guide are likely to be discussed in a session. This happens because, as participants express their ideas, experiences, and explanations on a subtopic, their comments are then supported, criticized, or elaborated on by others. Thus the few initial items often generate a fruitful chain of discussions not anticipated by the investigator. In a focus group interview on the use of contraceptives, for example, a woman may mention that although unmarried young women would prefer to use contraceptives, they do not want to purchase them at grocery stores because of social and religious taboos against premarital sex. This casual remark can lead to an interesting discussion of the mechanisms for distributing contraceptives. Participants are also likely to propose different ideas and suggestions that can be further examined by the group. The essential point is that the discussions in a focus group do not remain confined to the items in the interview guide. Thus, reliance on an elaborate interview guide can be self-defeating to the extent that it inhibits the free flow of ideas and insights. 3.2 Size and Composition of the Group The size of a focus group should not be so small that the advantages of group dynamics are lost. In small groups (3 or 4 participants), individuals feel exposed, and they feel a constant pressure to comment, whether they have something to say or not. Moreover, discussions in very small groups are more vulnerable to the effects of the personalities and opinions of influential participants. Conversely, a large group (more than 12 people) leaves little time for meaningful discussions. Individuals have to wait to make a comment, and by the time their turns come, they may have lost interest or forgotten the points they intended to make. As a result, spontaneity is undermined. Moreover, large groups tend to fragment into smaller subgroups, whose members address their remarks primarily to each other. The moderator may

19 be helpless in such situations. The optimal number of participants in a focus group interview is between 6 and 10 people. A group of this size is neither too large nor too small to permit the smooth flow of conversation. It is also easily manageable. To the extent possible, the focus group should be homogeneous in composition, with members sharing similar backgrounds and experience. For example, a focus group interview on the use of contraceptives should comprise people of generally the same age, gender, and socioeconomic background. This will require separate groups for men and women, old and young, and rich and poor. In the stratified societies of the developing world, participants drawn from different social and economic strata are unable to interact on an equal basis. Differences in status impinge on interpersonal communication. One generally finds that people of lower socioeconomic status are reluctant to talk in the presence of their perceived superiors. The presence of older women, for example, can deter unmarried girls from talking about their sexual behavior because they dare not disclose nontraditional behavior patterns in the presence of elders. In other instances, participants enjoying higher status might consciously or unconsciously dominate discussions. Only the valiant efforts of the moderator can save the situation. Efforts should be made to select people who do not know each other. Anonymity among participants also minimizes the inhibiting effect of differences in status and prevents the formation of small cliques, in which a few members talk primarily with one another and not with the whole group. Nonetheless, this requirement for anonymity cannot usually be met in rural areas where people have frequent contact with each other. If they do not personally know each other, they are still likely to know about each other. This is also the case when participants are drawn from an organization (e.g., extension service or family planning agency), because it is likely that they are familiar with the status and roles of the other members. The moderator should also try to exclude people who have previously participated in a focus group on the same subject. Repeat participants are not spontaneous in their responses and display a tendency to show off their past experience. Hence, what is much-prized previous experience in other circumstances is a disqualification for a focus group interview. Fortunately, the investigator can always find people who have not participated in focus group interviews because such interviews are uncommon in project or program areas. 3.3 Sampling Probability sampling to generate an unbiased sample

20 representative of the total population is not used to select participants for focus group interviews. This is neither possible nor desirable because of time and space constraints. The common accepted practice is to rely on convenience sampling, in which participants are selected on the basis of their easy availability, provided they meet some other predetermined criteria. However, efforts should still be made to see that participants are as representative of the target populations as possible. One can attain an acceptable degree of representativeness (1) by classifying the target populations on the basis of carefully selected criteria variables relevant to the study objective and (2) by including participants from each category in different groups. This can be explained with a simple illustration. Consider the owners of the tractors in an area development project. They can be classified on the basis of variables such as age (young or old), gender, size of holdings (smallholder or largeholders), and literacy (literate or illiterate). They can also be categorized according to use of the tractors (those who use tractors for their own farms, and those who rent them to others), source of financing (those who got loans from the project, and those who raised money by themselves), or the mechanical skills of the owners. The choice of a classification system will be dictated by the purpose of the inquiry. An investigator interested in discovering the cost-effectiveness of tractors for the owners might use categories based on size of landholding, source of financing, and the mechanical skills of the owners -- all factors that affect the cost-effectiveness of tractor operations. By contrast, if the purpose of the focus group interviews is to identify gender-based differences in the use of tractors by farmers, such elaborate categories are unnecessary -- a simple classification based on gender will be sufficient. Once appropriate categories have been identified, the focus groups can be formed in such a way that participants from all relevant categories are represented. For example, in focus group interviews for determining the cost-effectiveness of tractor operations, one can include (1) smallholders, (2) large landowners, (3) owners who cultivate their own farms, (4) those who rent, (5) owners who borrowed money to purchase tractors, (6) those who financed their purchase themselves, (7) those who possess some mechanical skills, and (8) those who do not. Efforts should be made to hold separate group sessions for each category or to include a mix of tractor owners from the various categories in each group session. To identify participants in developing countries, the best approach is to consult key informants who are knowledgeable about local conditions. The investigator can ask them to identify individuals who might be readily available for a focus group interview. It is always prudent to consult several informants to minimize the biases of individual preferences (see Section 4.2). Once the list is prepared, the investigator can select the